Consider this PostgreSQL session at REPEATABLE READ. Session B commits an INSERT of a matching row after session A's first SELECT. What does session A's second SELECT count, and why?

-- Session A (REPEATABLE READ), 10 rows have status='active' BEGIN ISOLATION LEVEL REPEATABLE READ; SELECT count(*) FROM jobs WHERE status='active'; -- returns 10 -- meanwhile Session B: INSERT INTO jobs(status) VALUES('active'); COMMIT; SELECT count(*) FROM jobs WHERE status='active'; -- returns ?